Professional Documents
Culture Documents
Webster R086 - JASAw
Webster R086 - JASAw
Stephen Bunting
Line array systems are now the forerunner in large scale live sound
reinforcement. This report describes why they have become so accepted and
widely-used over the last two decades, explores different arrangements of
arrays and why these are necessary, the benefits they have over more
traditional horizontally clustered systems in both directionality and high
frequency throw and how these are achieved. It also illustrates practical
advantages, disadvantages, cabinet design and techniques for setup,
configuration and touring.
1 Background
In the 60s and 70s, sound reinforcement systems used in popular music concerts were
often not sufficient to compete with the high levels of cheering, clapping and screaming from
the audience. It was necessary to begin using an increasing number of speaker cabinets to
provide the sound pressure level required, which were usually horizontally arrayed in clusters
with a ‘point and shoot’ philosophy and stacked on the side of stage [Webb 2003, 1]. There
were some inherent difficulties with this system.
• There was often an irregular frequency response due to destructive interference from
the closely stacked cabinets; also added to by the lack of HF coupling and the relative
abundance of LF coupling.
• Systems were generally short throw, again due to the lack of HF coupling, requiring
the use of delay stacks at regular intervals.
These systems started to be superseded by line array systems in the 80s and 90s. These
feature a vertical arrangement of specially designed loudspeakers that produce a highly
directional sound beam in the vertical plane. They were developed from earlier designs, also
known as column loudspeakers, which comprise a tall cabinet with a number of equally
spaced, identical drivers, and were historically installed in reverberant environments, such as
churches or railway stations to aid vocal clarity for announcements.
Among the advantages of line arrays are their increased directivity in the vertical plane
allowing HF to be projected further, a more consistent frequency response, improved direct to
reverberant ratio and the convenience of a modular package that can easily be flown above
the stage [Webb 2003, 1-2] [Klepper 1963, 198].
2001, 6-7].
Plotting this equation for various values of l (figure 2), shows that the distance to the far
field increases with increasing frequency (as stated by Eargle), and also with the length of
array. To take an example, for a 4m line array, we get transitions to far field at about 2m at
100Hz, 24m at 1kHz and 242m at 10kHz, while these are more like 10m at 100Hz, 100m at
1kHz and 1000m at 10kHz for an 8m long line.
The transition point has also been estimated by a similar procedure, but stating the
difference between r and r’ should be within a half-wavelength instead of a quarter-
wavelength [Kinsler 1982, 187-188].
There are a number of ways of estimating where the far field may start, but the most
important thing is to consider that it’s frequency dependant.
3 Directivity
Increased directivity in the vertical plane is a great advantage of line arrays, and is
covered by Mark Ureda in great detail in several AES publications, as well as in well-
documented and referenced acoustics theory from 70 years ago.
Plotting this for fixed n=2 and variable λd (figure 3) shows that there are severe lobes
produced when λd > 0.5 . Note that these lobes are of equal level to the on-axis lobe, and
Plotting the same variable for fixed n=5 (figure 4) shows the same effect, but more
pronounced.
d
λ = 0.25 d
λ = 0.5 d
λ = 0.75
d
λ =1 d
λ =2 d
λ =4
d
Figure 3 – Varying values of λ for 2 point sources arranged in a line.
d
λ =1 d
λ =2 d
λ =4
d
Figure 4 - Varying values of λ for 5 point sources arranged in a line.
Figure 5 - A straight line source with length L, with the path r(x) from point x on the line array to the
observation point P(r) (not shown)
As stated, considering the observation point, where the pressure is P(r), to be a great
distance away compared to the length of the array simplifies the equation, because if we were
to move x to a different point along the line source, r(x) changes insignificantly. Therefore:
1 1 1 1
≈ L ≈ − L ≈ (Eq. 3.3)
r ( x) r ( 2 ) r ( 2 ) r
However, the difference in distances between r(x) at different points along the array is
significant compared to a wavelength, therefore it must be taken into consideration. Figure 6
shows x repositioned along L, and the extra bit of distance that needs accounting for.
r ( x ) = x sin α (Eq. 3.4)
Figure 6 - x sin α represents the extra distance to a distant point from 2 parallel lines
P(r )
R(α ) = (Eq. 3.6)
Pmax
Pmax , the maximum radiated pressure, occurs when all segments along the line radiate in
phase. This is given as:
L
2
1
r −∫L
Pmax = A( x)dx (Eq. 3.7)
2
Therefore:
L
2
∫ A( x)e
− j ( kx sin α + Φ ( x ))
dx
−L
R(α ) = 2
L
(Eq. 3.8)
2
∫ A( x)dx
−L
2
As we are considering a uniform array, we can set A(x) to 1 (as amplitude is equal and
unity) and Φ (x) to 0 (as phase is equal), and we get the following:
L
2
∫e
− jkx sin α
R (α ) = dx (Eq. 3.9)
−L
2
To solve the integral gives the most commonly used expression for directivity of a line
array:
sin( kl2 sin α )
R(α ) = (Eq. 3.10)
2 sin α
kl
l
λ = 0.25 l
λ = 0.5 l
λ = 0.75 l
λ =1
l
λ = 1.5 l
λ =2 l
λ =3 l
λ =5
l
Figure 7 - The relationship between λ and the directivity of a uniform line source. This table shows that if the
array stays a constant length, the directivity pattern narrows and creates side lobes at higher frequencies.
Plotting this equation for changing ratios of λl , as in figure 7, shows that as the length of
the array increases, a constant frequency will get more directional. Also, as the wavelength
decreases (i.e. the frequency increases) for a given array length, the beam-width gets tighter
[Ureda 2001, 2-3].
Note also that the secondary lobes are significantly reduced from that of a multiple point
source model, and are considerably lower level than the primary lobe. This occurs because an
infinite amount of points can never be completely in phase, other than directly on-axis. The
effect is lessened at lower frequencies. Nonetheless, this does not mean that the lobes are
small enough to be negligible [Wolff 1930, 210-211].
The increased directionality for high frequencies is one of the main benefits of line array
systems. However, for a straight line array, this size can become far too narrow to be of any
use, as there will be a very slight position where the frequency response is just right, which is
far too small to cover a reasonable size audience.
and large angles for low λl . This is unsurprising as we’ve already shown that the directivity
beam gets narrower as the frequency gets higher.
If we consider that sin u ≈ u for small angles, equation 3.15 can be simplified to:
1.2λ
θ −6 dB = (Eq. 3.16)
l
Using degrees instead of radians and frequency instead of wavelength provides the
following approximate equation:
24000
θ −6 dB = (Eg. 3.17)
fl
This can be plotted to show frequency against quarter power angle with different array
lengths. You can see from figure 9 that at 10kHz, the -6dB point is only around 1° wide for
up to 8m long arrays and gets smaller for longer arrays. This is a very small angle, and is
practically unusable except for very specific applications [Ureda 2001, 4].
⎛σ ⎞ ⎛σ ⎞
rc (σ ) = 2 R sin ⎜ ⎟ sin ⎜ + α ⎟ (Eq. 3.18)
⎝2⎠ ⎝2 ⎠
We can then state the far field pressure at angle α as:
θ
2
∫θ A(σ )e
− j ( krc (σ ) + Φ (σ ))
P(r ) = dσ (Eq. 3.19)
−
2
∫θ A(σ )e
− j ( krc (σ ) + Φ (σ ))
dσ
−
Rcurved (α ) = 2
θ
(Eq. 3.20)
2
∫ A(σ )dσ
−θ
2
This can not be simplified any more as the straight line source was. [Ureda 2001, 3-4]. It
also unfortunately doesn’t lend itself easily to graph plotting. However, I have included plots
from Mark Ureda’s paper ‘J’ and ‘Spiral’ Arrays in figure 11, which clearly show that the
beam-width has increased for high values of λd , while the secondary maxima are still low.
Figure 11 – The polar plots of a curved source array for varying values of λd [Ureda 2001, 3-4]
where:
rL ( x) = x sin α (Eq. 3.23)
The equation for the circular portion is also the same, but the limits of integration must be
changed as it is rotated. So we get:
θ
Rcurved (α ) = ∫ e − jkrc (σ ) dσ (Eq. 3.24)
0
where:
⎛σ ⎞ ⎛σ ⎞
rc (σ ) = 2 R sin ⎜ ⎟ sin ⎜ + α ⎟ (Eq. 3.25)
⎝2⎠ ⎝2 ⎠
Again, this pattern does not lend itself easily to plotting, although if we compare the
pattern from a line, curved and J source of a 2 metre long array with a 60° curved source with
a radius of 1 metre (figure 13), the differences are quite clear. The J-Array has an asymmetric
plot, with a slightly narrower pattern than the curved source. Therefore, not only does a J-
array provide a more useful distribution pattern, it has a more practical directivity pattern.
Figure 13 – Response curves of a line source, curved source and J-array with line length 2m, curve of 60° and
radius of 1m [Ureda 2001, 4-5]
4.1 Planning
While line arrays would appear to be a step up from earlier sound reinforcement systems,
there is one big problem with them, which is that you can’t just stack your cabinets, aim them
in the general direction and expect them to perform well. There will have to be some
planning done beforehand to determine how many cabinets are needed, how to array them,
where they will be hung from, where they will be pointed and angle of curvature. This is a
greater problem on tour where you are presented with a new venue each day, rather than a
single installation [Bailey 2003].
Fortunately, most companies provide software that will estimate the characteristics of
different systems (manufacturer specific), such as JBL’s line array calculator (LAC) designed
for their VerTec range. This software helps the sound designer build the optimum rig based
on input of venue dimensions, seating planes, number of cabinets and it will display the
expected response from this input [JBL VerTec 4889 Owner’s Manual, Chapter 9].
This could help with planning a tour, as based on venue information received beforehand,
you can work out exactly what you need to take to each venue, and work everything out
before you get there.
Figure 14 – An example line array module with 2 LF drivers, 4 MF drivers and 3 HF drivers (example based on
JBL VerTec 4889).
5.1 Conclusion
The advantages of using high quality, carefully designed and implemented line array
systems over horizontal clusters and other types of speaker arrangement are many.
First of all, we can expect a higher perceived direct to reverberant ratio due to the
directional nature of the system. Of course, the reverberant level of the auditorium will still
be the same, but by directing as much sound energy as possible so as a member of the
audience will hear it before any reverb is desirable to achieve this effect. Preventing side
lobes which would direct sound towards the walls and create more reverberation also greatly
adds to this feature [Wolff 1930, 202-203].
Another consequence of the increased directionality is increased feedback rejection. With
correct microphone placement, which will usually be behind and to the side of the array, the
SPL radiated will be lower than with horizontally clustered sound reinforcement stacks
[Klepper 1963, 1].
Due to the increased HF throw of the system, the need for delay stacks is eliminated or
reduced (dependant on the venue). There is also a more even sound intensity over the entire
field, as the system doesn’t need to be worked exceptionally hard to get the level needed at
the back of the field. This is helpful for the front-of-house sound engineer who will get a
better impression of what’s happening in other areas of the venue.